Observation problems related to string vibrations. Outline of Ph.D. Thesis

Similar documents
6 Non-homogeneous Heat Problems

u tt = a 2 u xx u tt = a 2 (u xx + u yy )

Double Fourier series, generalized Lipschitz és Zygmund classes

Lucio Demeio Dipartimento di Ingegneria Industriale e delle Scienze Matematiche

Name: Math Homework Set # 5. March 12, 2010

Final: Solutions Math 118A, Fall 2013

MA 201: Method of Separation of Variables Finite Vibrating String Problem Lecture - 11 MA201(2016): PDE

Boundary-value Problems in Rectangular Coordinates

MATH-UA 263 Partial Differential Equations Recitation Summary

Math 220a - Fall 2002 Homework 6 Solutions

Wave Equation With Homogeneous Boundary Conditions

Partial differential equation for temperature u(x, t) in a heat conducting insulated rod along the x-axis is given by the Heat equation:

Spectrum and Exact Controllability of a Hybrid System of Elasticity.

G: Uniform Convergence of Fourier Series

MATH 819 FALL We considered solutions of this equation on the domain Ū, where

Lecture Notes on PDEs

(L, t) = 0, t > 0. (iii)

A proof for the full Fourier series on [ π, π] is given here.

Lecture 24. Scott Pauls 5/21/07

Applications of the Maximum Principle

(3) Let Y be a totally bounded subset of a metric space X. Then the closure Y of Y

Separation of Variables in Linear PDE: One-Dimensional Problems

Math 2930 Worksheet Wave Equation

Boundary value problems for partial differential equations

In this worksheet we will use the eigenfunction expansion to solve nonhomogeneous equation.

Section 12.6: Non-homogeneous Problems

Math Assignment 14

CHAPTER 10 NOTES DAVID SEAL

FOURIER SERIES, TRANSFORMS, AND BOUNDARY VALUE PROBLEMS

1 Math 241A-B Homework Problem List for F2015 and W2016

MA 201, Mathematics III, July-November 2016, Partial Differential Equations: 1D wave equation (contd.) and 1D heat conduction equation

MATH 131P: PRACTICE FINAL SOLUTIONS DECEMBER 12, 2012

u xx + u yy = 0. (5.1)

1 Wave Equation on Finite Interval

LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES)

Ordinary Differential Equations II

Diffusion on the half-line. The Dirichlet problem

MA 201: Differentiation and Integration of Fourier Series Applications of Fourier Series Lecture - 10

Partial Differential Equations

MA Chapter 10 practice

Math 4263 Homework Set 1

Separation of variables in two dimensions. Overview of method: Consider linear, homogeneous equation for u(v 1, v 2 )

Sufficient conditions for functions to form Riesz bases in L 2 and applications to nonlinear boundary-value problems

2.4 Eigenvalue problems

Partial Differential Equations Separation of Variables. 1 Partial Differential Equations and Operators

Controllability of linear PDEs (I): The wave equation

A CAUCHY PROBLEM OF SINE-GORDON EQUATIONS WITH NON-HOMOGENEOUS DIRICHLET BOUNDARY CONDITIONS 1. INTRODUCTION

ENGI 9420 Lecture Notes 8 - PDEs Page 8.01

A Guided Tour of the Wave Equation

Sturm-Liouville operators have form (given p(x) > 0, q(x)) + q(x), (notation means Lf = (pf ) + qf ) dx

Math 220A - Fall 2002 Homework 5 Solutions

Partial Differential Equations

MA 201: Partial Differential Equations D Alembert s Solution Lecture - 7 MA 201 (2016), PDE 1 / 20

Mathématiques appliquées (MATH0504-1) B. Dewals, Ch. Geuzaine

MATH 425, HOMEWORK 5, SOLUTIONS

Vibrating-string problem

Bilinear spatial control of the velocity term in a Kirchhoff plate equation

UNIVERSITY OF MANITOBA

MATHEMATICAL MODELS FOR SMALL DEFORMATIONS OF STRINGS

Mathematical Modeling using Partial Differential Equations (PDE s)

UNIVERSITY of LIMERICK OLLSCOIL LUIMNIGH

Math 342 Partial Differential Equations «Viktor Grigoryan

Chapter 3 Second Order Linear Equations

The lecture of 1/23/2013: WHY STURM-LIOUVILLE?

Partial Differential Equations

23 Elements of analytic ODE theory. Bessel s functions

Plot of temperature u versus x and t for the heat conduction problem of. ln(80/π) = 820 sec. τ = 2500 π 2. + xu t. = 0 3. u xx. + u xt 4.

Differential equations, comprehensive exam topics and sample questions

MATH 220: INNER PRODUCT SPACES, SYMMETRIC OPERATORS, ORTHOGONALITY

ENGI 9420 Lecture Notes 8 - PDEs Page 8.01

17 Source Problems for Heat and Wave IB- VPs

INTRODUCTION TO PDEs

About solving time dependent Schrodinger equation

My signature below certifies that I have complied with the University of Pennsylvania s Code of Academic Integrity in completing this exam.

Limits at Infinity. Horizontal Asymptotes. Definition (Limits at Infinity) Horizontal Asymptotes

(The) Three Linear Partial Differential Equations

ON THE CORRECT FORMULATION OF A MULTIDIMENSIONAL PROBLEM FOR STRICTLY HYPERBOLIC EQUATIONS OF HIGHER ORDER

On the spectrum of the Laplacian

Finite Difference Methods for Boundary Value Problems

1. Partial differential equations. Chapter 12: Partial Differential Equations. Examples. 2. The one-dimensional wave equation

Partial Differential Equations for Engineering Math 312, Fall 2012

MATH 251 Final Examination December 19, 2012 FORM A. Name: Student Number: Section:

THE WAVE EQUATION. F = T (x, t) j + T (x + x, t) j = T (sin(θ(x, t)) + sin(θ(x + x, t)))

Numerical computation of an optimal control problem with homogenization in one-dimensional case

1 Separation of Variables

3150 Review Problems for Final Exam. (1) Find the Fourier series of the 2π-periodic function whose values are given on [0, 2π) by cos(x) 0 x π f(x) =

Math 5587 Lecture 2. Jeff Calder. August 31, Initial/boundary conditions and well-posedness

Separation of Variables. A. Three Famous PDE s

SECTION (See Exercise 1 for verification when both boundary conditions are Robin.) The formal solution of problem 6.53 is

Notes on Fourier Series and Integrals Fourier Series

Wave Equation Modelling Solutions

UNIVERSITY of LIMERICK OLLSCOIL LUIMNIGH

Stabilization and Controllability for the Transmission Wave Equation

McGill University April 20, Advanced Calculus for Engineers

DETERMINATION OF AN UNKNOWN SOURCE TERM IN A SPACE-TIME FRACTIONAL DIFFUSION EQUATION

Fourier and Partial Differential Equations

Conservation Laws: Systematic Construction, Noether s Theorem, Applications, and Symbolic Computations.

Partial Differential Equations (PDEs)

MATH 590: Meshfree Methods

Electrodynamics PHY712. Lecture 3 Electrostatic potentials and fields. Reference: Chap. 1 in J. D. Jackson s textbook.

Transcription:

Observation problems related to string vibrations Outline of Ph.D. Thesis András Lajos Szijártó Supervisor: Dr. Ferenc Móricz Emeritus Professor Advisor: Dr. Jenő Hegedűs Associate Professor Doctoral School of Mathematics and Computer Science University of Szeged Faculty of Science and Informatics Bolyai Institute Szeged 016

1 Introduction The subject of the thesis is the solvability of observation problems related to vibrating strings, and the smoothness of their solutions. Observation problems have origins in control theory, which has a wide literature where attainability results were gained to various control conditions. We investigated observation problems in the following setting: the partial state of the vibrating string is known (observed) at two time instants, and we attempt to describe the whole vibration process, or at least determine an initial position and speed, from which the observed states are attained. The investigations of this type of problems is quite new, maybe the work [3] of L. N. Znamenskaya is the closest to our setting, where she studied observation problems related to the standard vibrating string with homogeneous boundary conditions. Further works similar to the subject of this thesis are [4], [5] and [6], which discuss observation problems related to vibrating beams, plates and membranes. In Chapter of the thesis, we investigate observation problems posed in terms of generalized functions. Our starting point is a vibration described by the Klein-Gordon equation, but later we managed to generalize our corresponding result with respect to the equation, to the boundary conditions and to the observed states. Chapter 3 generalize Duhamel s principle concerning the loaded infinite vibrating string. We gained this result in order to formulate the statements in the next chapter as sharp as possible, but it may be of interest also in itself. Chapter 4 focus on the observation problem of the infinite vibrating string utilizing classical tools. Our result can be extended to half-infinite strings with the help of the reflection method. The thesis is based on the [7] [10] publications of the author. We use the same notations and numbering (except for the references) in the outline as in the thesis. 1

Observation problems posed for vibrating strings in terms of generalized functions In this chapter of the dissertation, we use the definition of the spaces D s (S), s R given in [1]. Let the system {X n (x)} n=0 be a complete orthonormal basis in L (S). Given arbitrary real number s, we consider on the linear span D of the functions X n (x), n N 0 = N {0}, x S, the following Euclidean norm: ( ) 1 c n X n (x) := n s c n. s n=0 Completing D with respect to this norm, we obtain a Hilbert space denoted by D s. We use the notation S = (0, l) associated with string vibrations. Further n=0 information about the spaces D s can be found in []. The subject of Section.1 is a somewhat simple observation problem related to the Klein-Gordon equation. Our statement is the following: Theorem.1. Let the given observed states f 1, f and the observation instants t 1, t be such that f 1 D s+, f D s+, s R, and t t 1 = p q l, p, q N, a where p and q are relative primes. Moreover let us suppose, that sin ( (t t 1 ) Then the initial functions (nπ l a ) + c ) 0, n N. (ϕ, ψ) = (u(x, 0), u t (x, 0)) D s+1 D s

can be uniquely determined, such that the corresponding solution of the equation u tt (x, t) = a u xx (x, t) cu(x, t), (x, t) [0, l] R, 0 < a, c R with boundary conditions u(0, t) = 0, u(l, t) = 0, t R will satisfy the observation condition u(x, t 1 ) = f 1 (x), u(x, t ) = f (x), 0 x l. The proof is constructive, the Fourier series representations of the initial functions can be found in the proof. Similar statements can be made in the cases when the observation condition gives the speed of the string both at t 1 and at t ; or when the observation condition gives the position at one of the time instants, and the speed at the another. During our research, we managed to generalize this result in terms of each the equation describing the vibration, the boundary conditions and the observed states. This generalization can be found in Section., and it solves the following observation problem (which is the main result of Chapter ): Consider the following mixed problem: (.9) u tt = (p(x)u x ) x q(x)u Lu, (x, t) [0, l] R, 0 < p, q C ([0, l]), (.30) u t=0 = ϕ(x), u t t=0 = ψ(x), (.31) U i [u] U i (u x=0, u x=l, u x x=0, u x x=l ) = 0, i = 1,, 3

where U 1, U independent, self-adjoint linear expressions, and the functions u, ϕ, ψ are from the generalized function space D s. Let us suppose, that for every s R and for every (ϕ, ψ) D s+1 (0, l) D s (0, l), this mixed problem possesses the following good properties: (.33)!u solution and u C(D s+1, R) C 1 (D s, R) C (D s 1, R), and u can be written in the following form: (.34) u(x, t) = [α n cos (ω n t) + β n sin (ω n t)] X n (x), (x, t) [0, l] R, n=0 LX n = ω nx n, U i X n = 0, i = 1,. Our observation conditions are some known linear combination of the position and the speed of the string at the time instants t 1 and t : (.3) A 1 u t=t1 + B 1 u t t=t1 = f 1, A 1 + B 1 > 0, A u t=t + B u t t=t = f, A + B > 0, where the coefficients A 1, A, B 1, B and the functions f 1, f are given. Let (.35) f 1 D s+, f D s+, s R, and assume, that there are constants 0 < A Q, B R, 0 < M 1 R and a series C n R\{0} such that (.36) ω n (t t 1 ) + γ n δ n = nπa + B + C n, 0 < M 1 n < C n, n N and C n 0 as n, and (.37) sin(ω n (t t 1 ) + γ n δ n ) 0, n N 0. 4

We use Lemma.10 to estimate the values of the above sine function for n large enough: Lemma.10. Let the series r n such that 0 < M/n < r n 0 with a positive constant M and let x 0 > 0 a rational number. Then there exists a threshold number N for every fixed d R, such that sin (nπx 0 + d + r n ) > M, n > N. n This estimation will ensure, that the Fourier series of the initial functions will converge and will construate functions with the required smoothness. The angles γ n, δ n [0, π) in conditions (.36), (.37) are uniquely determined by the following relationships: sin γ n = sin δ n = A 1 B 1 ω, cos γ A 1 + B1 n = n, ω n A 1 + B1 ω n A B ω, cos δ A + B n = n. ω n A + B ω n Our statement (which is the main result of Chapter ) is the following: Theorem.11. The observation problem (.9) (.37) described above has unique solution (ϕ, ψ) D s+1 D s, that is, the initial functions can be uniquely determined from which the observed states described by condition (.3) are attained during the string vibration. In Section.3, we give three examples to illustrate the applicability of Theorem.11 in the case of the Klein-Gordon equation with various boundary conditions. We consider the vibrating string with fixed ends in Subsection.3.1, with free ends in Subsection.3., and with Sturm-Liouville boundary condition in Subsection.3.3. During these considerations we can see, that the rather technical conditions of Theorem.11 can be significantly simplified or they even automatically hold in certain cases. In Subsection.3.4 we investigate the case when we doesn t make any restrictions to the observation instants t 1 and t in Theorem.11. 5

3 A new version of Duhamel s principle for the infinite vibrating string In Chapter 3 of the thesis we introduce a new version of Duhamel s principle for the infinite vibrating string, which is the following: Theorem 3.1. If f(x, t) C(R ) and the directional derivative f t of f along t exists and f t C(R ), then the problem v(x, t) C (R ), v tt (x, t) a v xx (x, t) = f(x, t), (x, t) R, v t=0 = v t t=0 0 can be uniquely solved and the solution v can be written as v(x, t) = 1 t a 0 x+a(t τ) x a(t τ) f(ξ, τ)dξ dτ, (x, t) R. This theorem sharpens the smoothness condition f(x, t) C 1 (R ) of the classical result, and this representation of the solution v corresponds to the one in the usual setting. The base of the proof is that the function defined by this expression is well-defined and from C even with our weaker condition for the function f. Comment. We mention, that Theorem 3.1 (and accordingly the results of Chapter 4) remains valid, even if f(x, t) C(R ), f ν (x, t) := f ν C(R ), where f ν is the directional derivative of f along the vector ν = (ν 1, ν ), provided that ν is transversal to the characteristics. The proof of this statement 6

requires a tedious computation with several cases depending on the direction ν, hence we would like to communicate this proof in a subsequent work, which is in preparation. 4 Classically posed observation problems In Chapter 4, we use classical tools for investigating observation problems. To formulate as sharp statements as possible, we use the result of the previous chapter. The main result of the chapter is the following: Theorem 4.5. Consider the problem of the infinite vibrating string given by u(x, t) C (R ) u tt (x, t) a u xx (x, t) = f(x, t), (x, t) R, a > 0, where f(x, t) is continuous, and the directional derivative f t exists and f t C(R ). The observed partial states of the string are described by the conditions A 1 (x)u t=t1 + B 1 (x)u t t=t1 = f 1 (x), x R, A (x)u t=t + B (x)u t t=t = f (x), x R, where the given coefficients and right-hand sides satisfy A 1 (x), A (x), B 1 (x), B (x) 0, x R, A 1, A, B 1, B, f 1, f C 1 (R). This observation problem can be solved, namely there can be found initial functions u t=t0 (x) = ϕ(x), u t t=t0 (x) = ψ(x), ϕ C (R), ψ C 1 (R) such that the corresponding vibration described by u(x, t) produce the observed states. The solution is not unique. 7

This result can be extended to half-infinite strings with the help of the reflection method. Let us consider the equation of the half-infinite (x 0) vibrating string: (4.19) u(x, t) C ([0, ) R), (4.0) u tt (x, t) a u xx (x, t) = g(x, t), (x, t) [0, ) R, a > 0, where (4.1) g(x, t) C([0, ) R), g t (x, t) C([0, ) R). The endpoint is fixed, namely (4.) u(0, t) = 0, t R, and the observed partial states are the following: (4.3) u t=t1 = g 1 (x), u t=t = g (x), g 1, g C ([0, )). Theorem 4.3. The observation problem (4.19) (4.3) has a solution u for any functions g 1, g and g, which satisfy g(0, t) = g 1 (0) = g 1 (0) = g (0) = g (0) = 0, t R. The solution is not unique, but every solution u can be written in the following form: 8

u(x, t) = f 1(x a(t t 1 )) + f 1 (x + a(t t 1 )) + 1 a + 1 a t t 1 x+a(t τ) x a(t τ) f(ξ, τ)dξ dτ, x+a(t t 1) x a(t t 1) ψ(s)ds+ where the right-hand side is restricted to x 0. The functions f 1 and f are the odd extensions of the functions g 1 and g to the real numbers respect to the x = 0 point/line, and the function Ψ(x) = x 0 ψ(s)ds C (R) can be chosen arbitrarily inside the interval [0, T ]. In the case of the Neumann boundary condition (4.6) u x (0, t) = 0, t R, using even extensions we get the following results: Theorem 4.4. The observation problem (4.19), (4.0), (4.1), (4.3), (4.6) has a solution u for any functions g 1 and g, which satisfy g 1(0) = g (0) = 0. 9

The solution is not unique, but it can written as u(x, t) = f 1(x a(t t 1 )) + f 1 (x + a(t t 1 )) + 1 a t t 1 x+a(t τ) x a(t τ) + 1 a f(ξ, τ)dξ dτ, x+a(t t 1) x a(t t 1) ψ(s)ds, where the right-hand side is restricted to x 0. Here the function f 1 is the even extension of g 1, the function f(x, t) is the even extension of g(x, t) with respect to the x = 0 point/line. The function Ψ(x) = x 0 ψ(s)ds C (R) can be chosen arbitrary inside the interval [0, T ]. References [1] Komornik, V., Exact controllability and stabilization. The multiplier method, Research in Applied Mathematics 36, Chichester: Wiley, Paris: Masson, 1994. [] Lions, J. L., Magenes, E., Problèmes aux limites non homogenès et applications I-III, Dunod, Paris, 1968 1970. [3] Znamenskaya, L. N., State observability of elastic string vibrations under the boundary conditions of the first kind, Differential Equations, 46 (010), 748 75. [4] Egorov, A. I., On the observability of elastic vibration of a beam, Zh. Vychisl. Mat. Mat. Fiz. 48:6 (008), 967 973. 10

[5] Komornik, V., Loreti, P., Multiple-point internal observability of membranes and plates, Applicable Analysis 90:10 (011), 1545 1555. [6] Vest, A., Observation of vibrating systems at different time instants, E. J. Qualitative Theory of Differential Equations, 14 (014), 1 14. [7] Szijártó, A., Hegedűs, J., Observation problems posed for the Klein- Gordon equation, E. J. Qualitative Theory of Differential Equations, 7 (01), 1 13. [8] Szijártó, A., Hegedűs, J., Observability of string vibrations, E. J. Qualitative Theory of Differential Equations, 77 (013), 1 16. [9] Szijártó, A., Hegedűs, J., Classical solutions to observation problems for infinite strings under minimally smooth force, Acta Sci. Math., 81 (015), 503-56. [10] Szijártó, A., Hegedűs, J., Vibrating infinite string under general observation conditions and minimally smooth force, E. J. Qualitative Theory of Differential Equations, submitted 11